AI and prejudice: A science/fiction story based entirely on real events

Robert Elliot Smith

14:30

Business stage

We are all more connected than ever before, yet never more divided. We all have access to more information than ever, yet we can’t trust anything we read, hear, or see. We are all plugged into a more diverse world, yet the ugly intolerances of the past are resurgent. We all carry “smart” devices, yet arguably, the intelligence of society is retreating, not advancing.

Rob grew up in Alabama in the 1960s and 70s, then had a 30-year career in artificial intelligence (AI). To him, the online digital segregation that is emerging today looks a lot like the divided communities of his youth. Based on his experience with AI, he thinks that our “smart” devices are a big part of the problem, but that they could also be part of the solution, if we only reconsider from whence they came.

Rob’s research reveals that statistical ideas in the algorithms in all our lives today originated in the turn-of-the-20th-century social movement called eugenics. While initially a Darwin-inspired, progressive, utopian vision, eugenics was a wrong turn in science that became the basis of racist laws in America, and the worst atrocities of Nazi Germany. While completely deprecated as a social movement after WWII, eugenic ideas persist in statistical approaches to human social science. Today those approaches are embedded in the world's algorithmic infrastructure, the programs that effect all of us every day.

Ironically, eugenics originated at UCL, where Rob currently holds a post. So his story connects a personal journey, little-known scientific histories, and the computer programs that are running in everyone’s pockets today. Those connections provide insight not only into how we reached this frightening point in human social evolution, but how we can come out the other side with a new appreciation of humanity, its algorithmic inventions, and how the two can better work together.

Dr Robert Elliott Smith PhD FRSA is an expert in evolutionary algorithms, a researcher, and a practitioner in A.I. with 30 years’ experience. He has helped create software systems that learn fighter jet manoeuvres, describe immune system behaviour, reveal emotion in financial markets, and suggest how social networks propagate political polarisation. He has conducted research projects for Los Alamos and Oak Ridge National Laboratories, NASA, Boeing, NSF, EPSRC, the European Union, British Aerospace, Airbus, and British Telecom. He is the CTO of BOXARR Ltd., a fellow of the RSA, and a member of the Computer Science Faculty at UCL, where he helped found The Centre for Decision-Making Uncertainty.